Earned Value Management: CPI, SPI and Forecasting EAC

Most project reports tell you how much money is gone. That is a rear view mirror. It answers the question “how much have we spent” without answering the question that actually matters to a sponsor, which is “are we going to finish on budget and on time.” Earned Value Management, usually shortened to EVM, closes that gap. It adds one missing piece of information, the value of the work you have really completed, and that single addition lets you measure efficiency and forecast the finish while there is still runway to react.

This guide walks through the full EVM toolkit in plain language. You will see the four base numbers, the two efficiency indexes that summarize them, and the forecasting formulas that project the ending cost. We use one worked example the whole way through so the numbers stay connected, and we finish with the mistakes that trip up teams the first time they try it. If you would rather skip the arithmetic, the Earned Value Calculator (EVM) does every step for you, but reading the logic once makes the output far more useful.

What Earned Value Management measures

EVM rests on three measurements taken at a status date, which is just the day you stop and check the project. The first is Planned Value, or PV. This is the budgeted cost of the work that your schedule said should be done by now. If your plan called for 40 percent of the scope to be complete by today and the whole job is budgeted at 100,000 dollars, then PV is 40,000 dollars.

The second is Earned Value, or EV. This is the budgeted cost of the work that is actually done, measured at the original budget rate rather than at what it cost you. If you have genuinely completed 36 percent of the scope, then EV is 36,000 dollars, no matter what you paid to get there. The third is Actual Cost, or AC, which is exactly what it sounds like: the real money spent to complete that work.

Two more numbers set the frame. Budget at Completion, or BAC, is the total approved budget for the whole project. Estimate at Completion, or EAC, is your revised forecast of what the project will really cost once you account for how efficiently the team is working. PV, EV and AC are the inputs. BAC is fixed. EAC is what you calculate. Everything else in EVM is built from these ingredients.

The formulas, grouped so they make sense

There are more EVM formulas than most people want to memorize, so it helps to group them by the question each one answers. The variance formulas tell you where you stand right now. The index formulas turn those variances into ratios you can compare across projects. The forecasting formulas project the finish.

The reading rule is short. For CV and SV, negative is bad. For CPI and SPI, below 1.0 is bad. A CPI of 0.80 means you are getting 80 cents of value for every dollar you spend. An SPI of 0.90 means you have finished 90 percent of the work you planned to have finished by now. Anything above 1.0 means you are ahead on that dimension.

How to calculate step by step

Working the numbers in a fixed order keeps you from mixing up cost and schedule, which is the most common slip. Follow the same seven steps every time and the picture builds itself.

Start by writing down your three inputs, PV, EV and AC, plus BAC. Second, compute the two variances by subtracting: CV is EV minus AC, and SV is EV minus PV. Third, compute the two indexes by dividing: CPI is EV over AC, and SPI is EV over PV. Fourth, forecast the finish with EAC as BAC divided by CPI. Fifth, subtract to get VAC, which is BAC minus EAC. Sixth, subtract again to get ETC, which is EAC minus AC. Seventh, compute TCPI to see the efficiency you now need, using BAC minus EV over BAC minus AC.

Notice that steps two and three both start from the same inputs, and every later step reuses a number you already found. You never need a value out of thin air. If you can subtract and divide, you can run a full earned value analysis by hand.

The worked example, from inputs to forecast

Take a project budgeted at 100,000 dollars, so BAC is 100,000. At the status date the plan said 40,000 dollars of work should be done, so PV is 40,000. The work genuinely finished is worth 36,000 dollars at budget rate, so EV is 36,000. The team has spent 45,000 dollars to get there, so AC is 45,000. Those four numbers are everything we need.

Variances first. Cost Variance is EV minus AC, which is 36,000 minus 45,000, so CV is -9,000 dollars. The project is 9,000 dollars over budget for the work done. Schedule Variance is EV minus PV, which is 36,000 minus 40,000, so SV is -4,000 dollars. The project is behind plan by 4,000 dollars worth of work. Both are negative, so both stories are unhappy.

Indexes next. CPI is EV over AC, which is 36,000 divided by 45,000, so CPI is 0.80. You are earning 80 cents of value per dollar. SPI is EV over PV, which is 36,000 divided by 40,000, so SPI is 0.90. You have completed 90 percent of the planned pace. Now the forecast. EAC is BAC over CPI, which is 100,000 divided by 0.80, so EAC is 125,000 dollars. If the current cost efficiency holds, this project finishes at 125,000 rather than 100,000.

From there the rest falls out. VAC is BAC minus EAC, which is 100,000 minus 125,000, so VAC is -25,000 dollars, a forecast overrun of 25,000. ETC is EAC minus AC, which is 125,000 minus 45,000, so ETC is 80,000 dollars still to spend. Finally TCPI is BAC minus EV over BAC minus AC, which is 64,000 divided by 55,000, so TCPI is 1.16. To still finish inside the original 100,000, the team would need to run about 16 percent more efficiently than the plan for every remaining dollar, after already running under 1.0 so far. That is the honest signal that the budget target is at serious risk.

How to read and apply the result

Numbers are only useful when they change a decision, so treat the indexes as prompts for action. A CPI of 0.80 and an SPI of 0.90 together say the project is bleeding more on cost than on schedule. That points your attention at spend efficiency first: scope creep, rework, expensive resources, or an estimate that was optimistic from the start. If SPI had been the worse of the two, you would look at sequencing, dependencies, and whether the critical path is slipping.

The forecast is where EVM earns its keep with sponsors. Telling a steering committee “we have spent 45,000” invites a shrug. Telling them “at current efficiency this project lands at 125,000, which is 25 percent over budget, and here is the TCPI showing the recovery would take a 16 percent efficiency jump we have not yet demonstrated” turns a status update into a decision. They can add budget, cut scope, or accept the overrun with open eyes. EVM does not fix the project, but it makes the tradeoff visible while there is still time to choose.

One nuance on EAC. The formula BAC divided by CPI assumes the efficiency you have shown so far continues to the end. That is the standard forecast and it is usually the most defensible, because past performance predicts future performance better than hope does. But there are other EAC formulas for cases where you believe the overrun was a one time event, or where both cost and schedule pressure will compound. When the situation is more complex, the Estimate at Completion Calculator (EAC) runs all four standard forecast formulas side by side so you can compare the optimistic and pessimistic ends.

Common mistakes to avoid

The first mistake is confusing Earned Value with Actual Cost. EV is measured at the budget rate for work completed, not at what you paid. If you mix them, CPI becomes meaningless because you are dividing a number by itself. Keep the discipline that EV always uses the original budget, and AC always uses real spend.

The second mistake is faking percent complete. EVM is only as honest as the completion estimate feeding EV. If a task is reported at 90 percent for three weeks running, your EV is fiction and every index built on it lies to you. Use objective rules where you can, such as counting a deliverable as done only when it is accepted, rather than asking people how they feel about their progress.

The third mistake is reading SPI as if it were time. SPI is a value ratio, not a calendar. An SPI of 0.90 does not mean you are 10 percent behind in days, because tasks off the critical path can drag the index down while the finish date holds, and near the end of a project SPI drifts back toward 1.0 even when you deliver late. For schedule risk, pair SPI with a real look at the critical path using the Critical Path Calculator (CPM), and treat the two together rather than trusting either alone.

When EVM does not fit, and what to use instead

EVM shines when scope is defined, the budget is baselined, and progress can be measured against a plan. It struggles when those conditions are missing. On a discovery project where the scope is still forming, there is no meaningful PV to compare against, so the indexes wobble on shifting ground. On a small task of a few days, the setup cost of tracking PV, EV and AC outweighs the insight. And on pure time and materials work with no fixed deliverable, “percent complete” is hard to define honestly.

For those situations, reach for tools built to model uncertainty directly rather than force a plan onto it. When the estimate itself is the unknown, a three point estimate captures the range instead of pretending you have one number, and the PERT Calculator turns optimistic, likely and pessimistic guesses into an expected duration with variance. When you need to buy back time on a schedule that is already slipping, the Schedule Crashing Calculator finds the least cost way to compress the critical path. And when you are sizing a buffer for known risks before they hit, the Contingency Reserve Calculator (EMV) converts probability and impact into a defensible reserve. EVM tells you where you are. These tools help you decide what to do about it.

Three expert tips

Baseline before you measure, and freeze it

EVM compares reality to a plan, so the plan has to be stable. Set your budget and schedule baseline before work starts, then resist quietly editing it to match reality later. If scope genuinely changes, run a formal change so PV moves for a reason you can point to. A baseline that drifts every week gives you indexes that always read near 1.0 and warn you of nothing.

Trust CPI early, because it is stubborn

Research on completed projects has found that the CPI tends to stabilize by about the 20 percent complete mark and rarely improves on its own after that. When early CPI comes in at 0.80, do not assume the team will make it up later. Treat the early number as a forecast, model the finish with it, and act while you still have four fifths of the budget left to steer.

Report cost and schedule as a pair, never alone

A single index hides half the story. A project can be dead on budget while quietly slipping weeks, or dead on schedule while burning cash. Always show CPI and SPI together, and add the EAC forecast so the audience sees the ending rather than just the moment. The strongest one line status is a value like “CPI 0.80, SPI 0.90, forecast finish 125,000 against a 100,000 budget,” because it carries cost, schedule and outcome in a single breath.

Free project management calculators

Earned value is one tool in a connected set, and the numbers flow between them. Use the calculators below to run a full analysis without touching a spreadsheet, and start from the category hub if you want the whole toolkit in one place. Every tool below works in your browser and keeps your inputs private.

• Earned Value Calculator (EVM) – CV, SV, CPI, SPI and EAC from your PV, EV and AC.
• Estimate at Completion Calculator (EAC) – all four forecast formulas compared side by side.
• Critical Path Calculator (CPM) – forward and backward pass, float and the critical path.
• PERT Calculator – three point estimate, variance and completion probability.
• Schedule Crashing Calculator – the least cost time and cost tradeoff to compress a schedule.
• Contingency Reserve Calculator (EMV) – a risk based reserve from probability and impact.
• Project Management hub – the full set of planning and control calculators.

Frequently asked questions

What is the difference between Planned Value and Earned Value?

Planned Value is the budgeted cost of the work your schedule said should be done by the status date. Earned Value is the budgeted cost of the work that is actually done, measured at the same budget rate. PV is the target, EV is the achievement, and the gap between them is Schedule Variance.

What does a CPI below 1.0 mean?

It means you are over budget for the work completed. A CPI of 0.80 says you earn 80 cents of budgeted value for every dollar you spend. Any CPI under 1.0 signals cost inefficiency, and the lower it is, the larger the projected overrun.

How do I forecast the final cost of a project?

The standard forecast is Estimate at Completion, calculated as BAC divided by CPI. With a 100,000 dollar budget and a CPI of 0.80, EAC is 125,000. This assumes the current cost efficiency continues, which is usually the safest single assumption to report.

What is the difference between EAC and ETC?

Estimate at Completion is the forecast total cost for the whole project. Estimate to Complete is only the money still left to spend from the status date onward. ETC equals EAC minus Actual Cost. In the worked example, EAC is 125,000 and ETC is 80,000 because 45,000 is already spent.

What does TCPI tell me?

To Complete Performance Index is the cost efficiency you would need for the rest of the project to still finish inside the original budget. It is BAC minus EV over BAC minus AC. A TCPI of 1.16 means you must run 16 percent more efficiently than planned from now on, which is a warning when your CPI so far is only 0.80.

Is SPI the same as being behind schedule in days?

No. SPI is a ratio of work value, not calendar time. An SPI of 0.90 does not translate to being 10 percent late in days, because off critical path tasks can lower the index without moving the finish date. Use SPI as a directional signal and confirm real schedule risk with a critical path analysis.

Why is Earned Value measured at the budget rate instead of actual cost?

Because EV has to describe how much work is done, not how much you paid. If EV used actual cost, then CPI would always equal 1.0 and tell you nothing. Measuring EV at the original budget keeps it comparable to both Planned Value and Actual Cost, which is what makes the indexes meaningful.

When is CPI reliable enough to forecast with?

Studies of completed projects show CPI tends to settle by around the 20 percent complete point and rarely recovers much after that on its own. Once you are past that early stage, treat the CPI as a dependable input for your EAC forecast rather than assuming performance will improve.

Does EVM work for agile projects?

It can, with adaptation. Teams that use a stable backlog and a fixed release budget can define Earned Value from completed story points or accepted features. Where scope is still forming and there is no baselined plan, the indexes lose meaning, and range based tools such as PERT estimation fit the uncertainty better.

What inputs do I need to start an earned value analysis?

Four numbers: Budget at Completion, Planned Value, Earned Value and Actual Cost, all as of the same status date. From those you can derive every variance, index and forecast. The Earned Value Calculator asks for exactly these and returns CV, SV, CPI, SPI and EAC.

What should I do when CPI shows I am over budget?

First find the driver: rework, scope creep, or an optimistic estimate. Then present the EAC forecast and TCPI to the sponsor so the choice is explicit. The realistic options are adding budget, cutting scope, or accepting the overrun. Acting early, while most of the budget remains, gives you far more room than reacting at the end.

Earned Value Management turns a pile of spending data into a forecast you can defend. Learn the three inputs, keep the budget rate honest, and read cost and schedule together, and you will spot trouble while you can still steer around it. When you are ready to run the numbers on your own project, the Earned Value Calculator (EVM) and the wider Project Management toolkit are free and waiting.